Reinforcement Learning with Knowledge Representation and Reasoning: A Brief Survey